Method for non-destructive testing of regions of interest of a metal piece during additive manufacturing

The present invention relates to a method for non-destructive testing of regions of interest of metal pieces produced by additive manufacturing. The present invention also relates to an additive manufacturing system for non-destructive testing of regions of interest of a metal piece as well as a tangible computer product containing program code for causing a processor to execute the above method.

The recent evolution of the domain of additive manufacturing (AM) provides excellent tools not only for rapid and cost-effective manufacturing of prototypes and products but also for a single on-demand component, for a pre-production series of components and for production of a limited number of components. In particular, there is an increasing interest in manufacturing high value-added components with complex geometries that are hard or impossible to manufacture with traditional machining. However, the control of the quality of these components that have to comply with given quality standards and technical requirements so as to allow a use in highly demanding and regulated application fields is still a challenge.

In an additive manufacturing process, a manufacturing anomaly (e.g. non-intended variation of the structural or material properties of the final manufactured component) could arise due to a mismatch or an unintended drift of manufacturing parameters, an inconsistency in the 3-dimension (3D) model of the component, an incorrect application of a material layer, a flaw in the material powder, or a bad thermal management resulting in hot spots and related geometrical or material defects.

Depending on the relevance and of the typology of the variation as well as depending on the applied quality standards and technical requirements, the manufacturing anomaly can represent either an acceptable imperfection or an unacceptable flaw (i.e. defect) which could potentially lead to a malfunction or a premature mechanical failure of the component.

As AM stakeholders are striving to reach the same level of quality standards as traditional machining, there is a lack of relevant standards, guidelines, and recommendations for stakeholders to rely on. The immature nature of AM technologies causes uncertainties and increased risk exposure for involved stakeholders. Hence, the qualification of processes becomes complicated and time consuming for all involved parties.

Certifications of AM product rely on a manufacturing procedure qualification process. The qualification process ensures that the method by which parts are processed can meet the qualifying criteria in a repeated manner. The aim of the qualification process is to identify the variables of the process in order to know their influence on the part performance and process reproducibility, thereby providing a correlation between product and process specifications to ensure adequate and consistent performance of parts of AM products.

Various monitoring strategies are implemented by machine manufacturers to ensure the stability of the process. Among them, performing a layer-by-layer control of the build job using an NDT technique is an attractive approach to address the above challenges. This approach is described in several patents such as in WO2020/100022. A layer-wise surface scan with full coverage of the build area presents yet some drawbacks.

A layer-wise surface scan with full coverage of the build area is indeed costly and resource consuming in terms of power, data transfer rate, file size, etc. whereas most of the data of the scanned layers are not useful, as only a fraction of the layer information representing the manufactured cross-section part may comprise regions of interest that are more likely to present defects than the remaining cross-section part.

In addition, a full coverage scanning technology discloses the details of the internal geometry of the part to a third party that would review the NDT information which raises confidentiality issues with regard to the know-how of the manufacturer, thereby facilitating the unwanted reverse-engineering of the additive manufactured part by a competitor.

An aim of the present invention is, therefore, to provide a non-destructive method for certification of AM products that overcomes the shortcomings and limitations of the state of the art.

More particularly, an aim of the present invention is to provide a non-destructive testing method for certification of AM products or parts which keeps the part designed and the know-how of manufacturers confidential.

Another aim of the present invention is to provide a non-destructive testing method which advantageously streamlines the certification process of AM products.

A further aim of the present is to provide a method to provide a non-destructive method for monitoring AM products that is resource-efficient.

generating, on a computing device, a 3D digital model of at least one metal piece to be produced by the additive manufacturing system; identifying regions of interest in the 3D digital model where defects are more likely to occur during the additive manufacturing of said at least one metal piece. The regions of interest are identified using software configured to check the 3D digital model according to specific design rules. N NDT sensors are configured to be either: selectively activated among a total of M NDT sensors, or mounted on the sensor supporting structure at specific locations along said axis (X-axis). The N NDT sensors are selectively activated or mounted on said specific locations as a function of the regions of interest identified in the 3D digital model in order to retrieve data on the properties of the quality of metal parts of the metal piece only for the corresponding regions of interest of the metal piece during its additive manufacturing. According to the invention, these aims are achieved, according to an aspect of the invention, by a method of monitoring metal pieces produced by an additive manufacturing system. The system comprises a build platform, a material dispensing mechanism to dispense metallic powder on the build platform, and a sensor supporting structure configured to move across the build platform. N or M NDT sensors are mounted on the sensor supporting structure next to each other along an X-axis substantially perpendicular to the direction of the displacement of the supporting structure along a Y-axis. The method comprises the steps of:

In an embodiment, the N NDT sensors are selectively activated or mounted on said specific locations also as a function of the position of the sensor supporting structure on the build platform.

In an embodiment, the N NDT sensors are selectively activated or mounted on said specific locations also as a function of the position on the build platform where the at least one metal piece is to be manufactured and as a function of the build layer.

In an embodiment, the method further comprises generating at least one 3D digital model of a functionalized support supporting a part of the 3D digital model of the metal piece. The functional support comprises a plurality of control geometries having different build layers at different heights. The material property of each control geometry, once produced by additive manufacturing, is representative of the material property of a corresponding region of interest the metal piece supported by the functionalized support, said region of interest belonging to the same build layer of the corresponding control geometry.

In an embodiment, the functionalized support extends in the Z-direction during the additive manufacturing process. N NDT sensors are fixedly mounted on the supporting structure next to each other at a location along said axis to be aligned with the functionalized support in order to sense the control geometries during its additive manufacturing.

the identified regions of interest, the position of the sensor supporting structure on the build platform, and the build layer of the metal piece being manufactured. In an embodiment, the manufacturing system further comprises a multiplexer configured to selectively activate N NDT sensors among said M NDT sensors mounted on the sensor supporting structure, and a control unit configured to control the multiplexer, during the additive manufacturing of the metal piece, as a function of:

In an embodiment, M is at least 100 and N is comprised between 2 and 32, preferably between 4 and 16 and N varies for each layer or for some layers constituting the metal piece.

In an embodiment, the control unit comprises a processor. The 3D digital model of the metal piece to be produced is converted into a build job file comprising data of identified regions of interest in the form of program codes. The build job file is sent from the computing device to the control unit to be executed by the processor for causing the control unit to control the multiplexer as a function of the identified regions of interest and as a function of the position of the sensor supporting structure on the build platform.

In an embodiment, the additive manufacturing system further comprises a calibration standard positioned in the manufacturing system such that said sensor supporting structure may move above the calibration standard. The method further comprises moving the sensor supporting structure one or more times during the additive manufacturing of the at least one metal piece to retrieve one or more calibration data for each of the N NDT sensors.

continuously extend(s) along a direction parallel to the X-axis of the sensor supporting structure, or are distinct and identical calibration features positioned along said direction in correspondence with the position of the N NDT sensor.Each of the N NDT sensors of a given type of the sensor supporting support may therefore sense identical characteristics when the sensor supporting support moves across the calibration standard. In an embodiment, the calibration standard comprises one or more calibration features adapted to the type of NDT sensors used. The one or more calibration features either:

In an embodiment, the one or more calibration features comprise at least a first and a second portion having respectively a first and a second conductivity that are different.

In an embodiment, the one or more calibration features comprise machined features with at least one characteristic dimension below 0.2 mm, such as a notch, a step, or a hole.

moving the recoater one or more times above the calibration standard during the formation of one or more powder layers on the build platform, or moving the recoater one or more times beyond the build platform above the calibration standard after the formation of one or more corresponding layers. One or more calibration data may therefore be retrieved for each of the N NTD sensors. In an embodiment, the additive manufacturing system further comprises a material dispensing platform comprising a bed of metallic powder. The sensor supporting structure is a recoater configured to move the metal powder from the material dispensing platform over the build platform layer by layer. The calibration standard is positioned in the manufacturing system such that the recoater may move above the calibration standard. The method further comprises, during the additive manufacturing of the metal piece, the following step:

moving the recoater further along the Y-axis across the discharge opening to remove the excess of metallic powder for each additional layer spreads across the build platform and, moving the recoater even further, beyond the build platform, one or more times along the Y-axis to position the NDT sensors above the calibration standard during the additive manufacturing of the at least one metal piece in order to retrieve one or more calibration data for each of the N NTD sensors. In an embodiment, the additive manufacturing system further comprises a discharge opening positioned between the build platform and the calibration standard. The method comprises

In an embodiment, the calibration standard is either forming an integral part with the build platform or removably included in the build platform. The top surface of the calibration standard is substantially coplanar with the build platform surface.

In an embodiment, N NDT sensors are eddy current sensors.

a build platform actuable along a vertical axis and on which the at least one metal piece is to be manufactured, a material dispensing mechanism for dispensing metallic powder on the build platform, and a sensor supporting structure configured to move across the build platform. N or M NDT sensors are mounted on the supporting structure next to each other along an axis substantially perpendicular to the direction of the displacement of the sensor supporting structure. N NDT sensors are configured to retrieve data on the properties of the quality of metal parts of the metal piece corresponding of the regions of interest identified by the method as described above. Another aspect of the invention relates to an additive manufacturing system for additive manufacturing of at least one metal piece and for monitoring the metal piece. The system comprises:

In an embodiment, the system further comprises a material dispensing platform comprising a bed of metallic powder. The sensor supporting structure is a recoater configured to move the metallic powder from the material dispensing platform over the build platform layer by layer.

In an embodiment, the system further comprises a discharge opening for discharging the excess of metallic powder and a calibration standard. The discharge opening is positioned between the build platform and the calibration standard such that the recoater is arranged to move across the discharge opening to remove the excess of metallic powder and then above the calibration standard.

In an embodiment, M is at least 100 NDT sensors arranged in the recoater. The system further comprises a multiplexer configured to selectively activate N NDT sensors among said at least 100 NDT sensors, a control unit configured to control the multiplexer and a measurements unit connected to the multiplexer to retrieved data on the property of the quality of metal parts of the at least one metal piece corresponding to the selected regions of interest of the at least one metal piece.

In an embodiment, the M NDT sensors are eddy-current sensors.

Another aspect of the invention relates to a tangible computer product containing a build job file in the form of program codes to be executed by the processor of the method as described above.

generating, on a computing device, a 3D digital model of the build volume and a 3D digital model, in the build volume, of at least one metal piece to be produced by the additive manufacturing system; 3 selecting empty regions in the build volume, wherein each empty region has a volume larger than 10×10×10 mm, and generating 3D digital models of control geometries in the empty regions. N NDT sensors are configured to be either selectively activated among a total of M NDT sensors, or mounted on the sensor supporting structure at specific locations along said axis. The N NDT sensors are selectively activated or mounted on said specific locations as a function of the position of the 3D digital models of control geometries in the corresponding empty regions of the 3D digital model of the build volume and as a function of the position on the build platform where the control geometries are to be manufactured. Data on the properties of the quality of metal parts of the control geometries may therefore be retrieved during their additive manufacturing without directly sensing the at least one metal piece during its additive manufacturing. Another aspect of the invention relates to a method of monitoring metal pieces produced by an additive manufacturing system comprising a build platform, a material dispensing mechanism to dispense metallic powder on the build platform, a sensor supporting structure configured to move across the build platform, and a build volume. N or M NDT sensors are mounted on the supporting structure next to each other along an axis substantially perpendicular to the direction of the displacement of the sensor supporting structure. The method comprises the steps of:

In an embodiment, at least one of the 3D models of control geometries comprises a closed conductive portion surrounding an empty space. The closed conductive portion may have a regular width within a building plane. The thickness of said width is equal or thinner than the thickness of the thinnest portion of the 3D digital model of the at least one metal piece.

In an embodiment, the NTD sensors are eddy current sensors comprising each a coil generating a time-varying electromagnetic field within the range from 1 kHz to 10 MHz. The coil's sensing size characteristics are substantially identical to the dimensions of the closed conductive portion of the control geometry when the coil is oriented in the building direction of the control geometry.

In an embodiment, the 3D digital models of control geometries have a maximum thickness of one millimetre and the corresponding manufactured control geometries are made of one or more layers selected from 1 to 30.

In an embodiment, each control geometry has at least one support structure linking said control geometry to the build plate. The support structure has on average a total sectional area in the build direction of no more than 30% of that of the reference geometry.

generating, on a computing device, a 3D digital model of the build volume comprising at least one metal piece to be produced by the additive manufacturing system and one or more support structures supporting one or more parts of the at least one metal piece, and including in at least one support structure at least one control geometry. N NDT sensors are configured to be either selectively activated among a total of M NDT sensors, or mounted on the sensor supporting structure at specific locations along said axis. The N NDT sensors are selectively activated or mounted on said specific locations as a function of the location of the control geometries in one or more of the support structures in the 3D digital model of the build volume in order to retrieve data on the properties of the quality of the control geometries included in at least one support structure during its manufacturing. Another aspect of the invention relates to a method of monitoring metal pieces produced by an additive manufacturing system comprising a build platform, a material dispensing mechanism to dispense metallic powder on the build platform, a sensor supporting structure configured to move across the build platform, and a build volume. N or M NDT sensors are mounted on the supporting structure next to each other along an axis substantially perpendicular to the direction of the displacement of the sensor supporting structure. The method comprising the steps of:

In an embodiment, the N NDT sensors are selectively activated or mounted on said specific locations also as a function of the position of the sensor supporting structure on the build platform.

generating, on a computing device, a 3D digital model of the build volume comprising at least one metal piece to be produced by the additive manufacturing system, identifying regions of interest in the 3D digital model of the at least one metal piece where defects are more likely to occur and/or are more critical during the additive manufacturing of said at least one metal piece, wherein the regions of interest are identified using software configured to check the 3D digital model according to specific design rules, and modifying in the 3D digital model of the build volume the position and/or orientation of the 3D digital model of the at least one metal piece so as to maximize the number of regions of interest that are located on an axis parallel to the X-axis of the sensor supporting structure before starting the additive manufacturing of the at least one metal piece. N NDT sensors are configured to be either selectively activated among a total of M NDT sensors, or mounted on the sensor supporting structure at specific locations along said axis. Data of the properties on the quality of metal parts may therefore be retrieved for an optimized number of corresponding regions of interest of the metal piece during its additive manufacturing. Another aspect of the invention relates to a method of monitoring metal pieces produced by an additive manufacturing system comprising a build platform, a material dispensing mechanism to dispense metallic powder on the build platform, a sensor supporting structure configured to move across the build platform, and a build volume. NDT sensors are mounted on the supporting structure along an axis substantially perpendicular to the direction of the displacement of the sensor supporting structure. The method comprises the steps of:

Another aspect of the invention relates to a method of monitoring metal pieces produced by an additive manufacturing system comprising a build platform, a material dispensing mechanism to dispense metallic powder on the build platform, a sensor supporting structure configured to move across the build platform, and at least one NDT sensor affixed to the sensor supporting structure. The NDT sensor moves above a calibration standard several times during the additive manufacturing of a metal piece in order to retrieve a calibration dataset associated with an additive manufacturing run.

In an embodiment, the height of the top surface of the calibration standard is within +/−5 mm of the Z position of the plane corresponding to the upper surface layer formed by the material dispensing mechanism.

In an embodiment, the calibration standard is either forming an integral part with the build platform or removably included in the build platform. The top surface of the calibration standard is substantially coplanar with the build platform surface.

In an embodiment, the calibration standard is marked with a unique identifier that is visible and/or readable through electromagnetic means.

In an embodiment, the NDT sensor is an optical sensor.

In an embodiment, the NDT sensor is an eddy current sensor.

In an embodiment, the calibration standard comprises regions with different conductivities.

In an embodiment, the calibration standard comprises machined features with at least on characteristic dimension below 0.2 mm, such as a notch, a step, or a hole.

In an embodiment, a powder evacuation opening or recess is located between the build platform and the calibration standard.

In the context of the present invention, a metal piece shall be understood as a piece substantially made of metal as such piece may be made of metallic powder either sintered/welded using the SLM technology or held together by a non-metallic binding agent using the binder jetting technology.

In addition, specific design rules in the context of the present invention shall be understood as rules providing direct guidelines or constraints in relation to part designs, process parameters, and material properties. AM technologies provide tremendous flexibility for designers because of the wide range of complex geometries that they can produce. For this reason, design rules in AM are desired because they can provide much-needed insight into manufacturability for a particular material, design, or process. That insight leads to an understanding of direct guidelines or constraints during process planning. Design rules can be critical to satisfying manufacturability because they enable designers to determine the best process and material combinations for a design. The skilled person knows how to select design rules as a function of the geometry of the metal piece to be manufactured and the process parameters associated with a particular additive manufacturing system.

Further, in the context of the present invention, a build job file may include not only data on the shape of the metal piece to be produced but also, data pertaining to the position/orientation of the metal piece on the build platform. Real-time information on the relative position between the metal piece being produced and the sensor supporting structure may however require additional data to be fed to the controller of the supporting structure of the additive manufacturing system to take into consideration non-constant travel speed of the sensor supporting structure as described subsequently.

Furthermore, a build layer as described therein shall be understood as a layer index, i.e. a specific layer of multiple layers constituting the metal piece produced by the additive manufacturing process.

1 FIG. 100 10 Referring to, additive manufacturing of a metal pieceis produced by a Selective Laser Melting (SLM) systemaccording to an embodiment of the invention. This technique can be used for the additive manufacturing of stainless steel, tool steel, titanium, cobalt chrome, and aluminum parts. Other additive manufacturing systems based on other AM technologies may be used within the context of the present invention such as binder jet 3D printing systems.

10 20 22 20 24 16 20 24 24 16 The SLM machinecomprises a material dispensing platformconfigured to move upwardly during additive manufacturing, a metal powder bedon the dispensing platform, a build platformconfigured to move downwardly during additive manufacturing and a recoaterconfigured to move metal powder layer by layer from the dispensing platformto the build platform. There are however other solutions to dispense the metal powder on the build platform, for example by a container mounted on an upper portion of the recoater and adapted to dispense metal powder on the build platformas the recoatermoves across the platform.

10 12 14 50 100 50 3 FIG. 5 FIG. The SLM machinecomprises a laser sourceand a mirror scannerconfigured to move laser beams over each layer of metal powder to fuse portions of said layer of metal powder as a function of a 3D digital model, as shown in, previously generated by an AM design software program in order to produce a metal pieceaccording to the 3D digital modelas illustrated in.

50 10 100 AM design software packages usually comprise a module that can be used to perform numerical simulation using the 3D digital model of the metal piece to be manufactured and manufacturing parameters. The software is therefore capable of simulating for example mechanical stresses as wells as thermal stresses on the 3D digital modelof a specific shape as a function of the design rules associated with this specific shape and with the parameters of the additive manufacturing systemin order to identify regions of interest that are more likely to present defects which could affect the overall quality of the metal piece. More particularly, the region of interest may be identified as a function of the response to mechanical and/or thermal stresses exceeding a given threshold. The threshold may be determined according to the field of application of the metal piece.

For example, the threshold may be set at a low value for metal pieces intended to be used in a combustion engine as there are subject to very high mechanical and thermal stresses. If the response affects the integrity of the metal piece to a degree that could potentially lead to a malfunction or a premature mechanical failure of the metal piece, the 3D digital model is adapted to provide a more robust metal piece capable to withstand high mechanical and thermal stresses in the long-term.

2 FIG. 1 FIG. 18 16 18 16 18 18 Referring to, which illustrates an advantageous embodiment, M NDT sensorsare mounted on the recoaternext to each other along an axis (X-axis) perpendicular to the direction of displacement of the recoater (Y-axis) as shown in. The number of NDT sensorson the recoateris preferably above 50, preferably above 80, even more preferably above 100. In another non-illustrated embodiment, the NDT sensorsmay be mounted on a sensor supporting structure which is independent of the recoater. The sensorsare arranged next to each other along an axis (Y-axis) substantially perpendicular to the direction of the displacement of the sensor supporting structure across the build platform.

18 The NDT sensorsmay be selected for example among: an eddy current sensor, a temperature-sensitive sensor, a capacitive sensor, an optical sensor comprising an optical sensing element sensitive to visible and/or infrared (IR) and/or ultraviolet (UV) radiations, and a laser sensing system able to generate and detect ultrasonic waves.

18 16 16 24 18 18 18 In an advantageous embodiment, M eddy current sensorsare mounted on the recoaternext to each other, as described above, in order to be able to sense newly solidified cross-section of the metal piece being manufactured as the recoatermoves across the build platformto spread another layer of metallic powder. Each eddy current sensormay be configured to generate an excitation magnetic field with a single, given excitation frequency. The eddy current sensoris configured to create a time-varying, local magnetic field and to measure the induced magnetic field response, or field variations response in a localized manner. The material properties of a portion of a newly solidified cross-section, which is sensed by the eddy current sensorduring the additive manufacturing of a metal piece, are a function of the electromagnetic properties of the sensed portion, such as its electrical conductivity and magnetic permeability, that depend on the integrity and of the particular shape of the sensed portion (absence of crack or non-uniformity within the solidified mass).

18 An eddy current sensing array, made of multiple eddy current sensors, can be configured in an absolute or send receive mode. A plurality of interrogating frequencies can be used for each coil to better separate surface and sub-surface defects. A plurality of geometries, such as larger (>3 mm) and smaller coils (0.5 mm<×<3 mm), can be used to increase the magnetic field penetration or the local resolution.

18 100 18 16 100 Each eddy current sensor, or each of any other type of NDT sensors mentioned above may sense a portion of a cross-section of the metal piecein the course of its manufacturing process. The sensed portion may be less than 3 mm×3 mm×PD, and advantageously less than 1 mm×1 mm×PD, wherein PD is the electromagnetic penetration depth (“skin” depth). The arrangement of the NDT sensorson the recoatermay therefore sense, in high resolution, entire cross-sections of the metal pieceto be inspected.

100 However, as mentioned earlier, sensing entire cross-sections of the metal pieceis not desirable as not only it raises confidentially issues on the know-how of the manufacturer without having necessarily more accurate information on the integrity of the metal piece, but it is also resource-consuming in terms of power, data transfer rate, file size, etc.

2 FIG. 5 FIG. 10 27 28 27 30 28 28 18 28 30 102 100 To address these issues, with reference to, the additive manufacturing systemfurther comprises a control unit, a multiplexerconnected to the control unitand a measurements unitconnected to the multiplexer. M outputs of the multiplexerare electrically connected to respective M NDT sensorswhile N inputs of the multiplexerare electrically connected to the measurement unitto retrieved data on the material property of the selected regions of interestof the metal pieceas shown in.

27 27 40 200 45 27 28 18 52 16 a 17 FIG. The control unitcomprises a processorconfigured to execute a build job file sent by a remote computing devicethrough the internetor by an in-situ computing devicelocated nearby the additive manufacturing system () in order to cause the control unitto control the multiplexerfor selectively activating N sensorsas a function of the selected regions of interestand the position of the recoateron the Y axis and optionally on the X axis.

27 28 102 16 100 10 16 27 18 The control unitof the additive manufacturing system therefore constantly controls the multiplexeras a function of both the regions of interestto be inspected and the position of the recoateras it moves along the Y-direction and possibly along the X-direction during the additive manufacturing process of the metal piece. In this respect, the additive manufacturing systemmay comprise sensors positioned to measure in real-time the position or the travel speed of the recoater. Position or speed data may be fed to the control unitto control the multiplexer as to ensure that each sensoris timely activated to sense the intended region of interest of the metal piece.

3 5 FIGS.to 3 FIG. 100 50 52 shows an example of the additive manufacturing process of a metal petal piece. In particular,shows the 3D modelof the metal piece to be produced. The 3D model comprises two lateral parts with cantilevered portions that have been selected as regions of interestas they are more likely to present defect by the nature of their respective shape.

4 FIG. 27 10 28 27 18 52 a Referring to, at the start of the additive manufacturing of the metal piece, the control unitof the additive manufacturing systemcontrol the multiplexerbased on the build job file executed by the processorso as to selectively activate N NDT sensorswhich are aligned with the base of the selected regions of interest.

100 27 28 102 52 5 FIG. As the additive manufacturing of the metal pieceprogresses as shown in, the control unitconstantly control the multiplexerto activate and deactivate the corresponding NDT sensors to ensure that only the regions of interestof the metal piece, corresponding to the regions of interestof the 3D digital model, are inspected.

18 100 16 The signal sensed by each NDT sensorfor the identified regions of interest of the metal pieceduring its additive manufacturing are processed as a function of the movement of the NDT sensors, which correspond to the movement of the recoaterwhen the NDT sensors are mounted thereon according to an embodiment.

0 1 In particular, time difference, for each build layer and for each sensor, between the instant tcorresponding to the beginning of the displacement of each NDT sensor in the Y direction and the instant tcorresponding to the instant when a sensor senses a region of interest is timestamp. Timestamp for each measurement of regions of interest are then processed as a function of position data which can be obtained either from the motion controller of the recoater or from a dedicated device in order to assign for each region of interest the corresponding measurement.

52 In another non-illustrated embodiment, N NDT sensors are mounted on the recoater, or on a sensor supporting structure independent of the recoater, next to each other along an axis (X-axis) substantially perpendicular to the direction of the displacement of the recoater (Y-axis) or along an axis (X-axis) substantially perpendicular to the direction of the displacement of the supporting structure (Y-axis). Each NDT sensor is positioned at a specific location on the recoater or on the sensor supporting structure as a function of the regions of interestidentified in the 3D digital model as described above.

102 60 6 FIG. It may however be difficult to measure the material property with NDT sensors for a metal piece comprising specific parts whose defects are poorly measured by NDT techniques, such as thin walls or lattice-structure. For metal pieces comprising parts that need to be supported during the additive manufacturing, such as cantilevered parts, a functionalized supportmay be used as illustrated inaccording to an embodiment.

18 16 60 70 70 70 70 70 70 60 60 60 a b c d e f In this embodiment, NDT sensorsare fixedly positioned on the recoaterto be aligned with the functionalized support. These sensors may sense only a cross-section of control geometries,,,,,integrated in the functionalized supportat different build layers at different heights along the Z-direction during the inspection of the functionalized support. Only a few NDT sensors are required to inspect the supportand may be comprised for example between 2 and 6.

10 102 102 102 102 102 102 100 a b c d e f 6 FIG. The material properties of each of these control geometries, once produced by the additive manufacturing system, are representative of the material properties of regions of interest,,,,,of the metal piecehaving the same respective build layer as shown in. The different control geometries of the functionalized support can therefore be used as reference geometries to test the quality of different portions of the metal piece at defined heights, to compare different metal pieces with the same build job files or even for metal pieces having different shapes built with the same material.

100 102 30 45 40 17 FIG. Once the additive manufacturing of the metal pieceis completed, data on the material property of the regions of interestacquired by the measurement unitare sent to the in-situ computing deviceor the remote computing devicefor analyses as shown in.

7 14 FIGS.to 70 100 24 80 70 100 Referring to, control geometriesthat are distinct from one or more metal piecesmay also be designed to be manufactured on the build platformin empty regionswhich can be selected in the design software in the build volume as a function of the position and/or orientation and the shape of one or more 3D digital models of one or more corresponding metal pieces on the build volume. Different parameters may be measured on the control geometriesand/or on the metal piece, each corresponding to a defect category such as: surface irregularities, e.g. projections or recesses which are typically measured by optical means; surplus of partly fused material; altered colour, especially for titanium, and/or traces of powder as a sign of insufficient atmosphere protection, which may be measured by optical means; porosity, cracking, lack of fusion and residual stress which can be measured in-situ by NDT sensors, geometric dimension inaccuracy and local displacement or warping due to internal stress which may be measured in-situ with NDT sensors; local conductivity and conductivity variations, which may be measured in-situ with NDT sensors; differences in the microstructure, for instance between fine equiaxed grain and larger columnar grain structures that can be measured with NDT sensors.

26 16 24 26 25 24 16 25 22 25 18 26 100 1 FIG. In an embodiment, a calibration standardmay be included on the trajectory of the recoateror of the sensor supporting structure. The calibration standard is placed in an X, Y-position in a recess on the build platformor on a recessbeyond a discharge openingpositioned next to the build platform, as shown in. The recoateris configured to further move along the Y-axis over the discharge openingto remove the excess of metallic powderfor each additional layer. The recoater is further configured to move further along the Y-axis beyond the discharge openingfor one or more layers to position the NDT sensorsabove the calibration standardduring the additive manufacturing of the metal piecein order to retrieve one or more calibration data for one or more NTD sensors. The calibration standard, therefore, allows for traceable calibration of the NDT sensors during each build job.

The term “calibration standard” should be interpreted in a broad sense and may refer to a reference block, a calibration block, a reference standard, etc.

26 18 16 26 18 16 The calibration standardcomprises one or more calibration features adapted to the type of NDT sensor used. The calibration feature should continuously extend along the X-axis such that the NDT sensorsof a given type of the sensor supporting structuremay sense identical characteristics when the latter moves across the calibration standardat least once during an additive manufacturing run. Alternatively, several identical calibration features should be located on the path of each NDT sensorof the sensor supporting structure.

26 The calibration standardmay advantageously comprise a marking encoding a unique identifier for traceability purposes. The marking is unique in the sense that calibration standards have an identifier allocated uniquely thereto that is sufficient to identify each in a non-ambiguous way. Each calibration standard can therefore be associated with a quality certificate mentioning said unique identifier that proves its fitness for purpose as means of calibration according to a given testing procedure. The marking may be visible and/or readable through electromagnetic means.

26 24 In an embodiment, the calibration standardis positioned outside of the build platform. It may for example be positioned beyond the build platform in the powder deposition direction. In this embodiment, the calibration standard has an upper surface substantially positioned in a plane coplanar with the plane of the upper surface layer of the metallic powder, i.e., within +/−5 mm in the Z direction of the plane of this upper surface layer, and preferably within +/−1 mm of this plane.

24 24 24 a a The calibration standard may be placed on the build platformsuch that the calibration top surface is substantially at the same height as the build platform surface. In other words, the calibration standard top surface is substantially coplanar with the build platform surface. The height of the top surface of the calibration standard above the build platform surface should in anyway not exceed 1 mm, and should typically be lower than 0.1 mm.

24 24 In an embodiment, calibration features are machined directly into the material of the build platform. An advantageous machined calibration feature may be a notch with a typical depth of 0.1 to 5 mm created by wire electrical discharge machining (wire EDM) across the top surface of the build platform. A machined calibration feature may be advantageously of constant depth in the Z direction, and in an orientation in the X, Y plane such that all NDT sensors can move over it within the travel of the sensor supporting structure.

24 In another embodiment, a calibration standard is positioned in a recess or hole of the build platformwith the advantage of having the choice of materials that may be used to realize the calibration standard. The calibration standard may also be removably attached to the build platform to facilitate the removal of printed metal parts and/or to perform maintenance operations. The calibration standard may be exchanged according to the quality standard used to control a given build job or shared between several build platforms.

A calibration measurement (i.e. a measurement performed when the NDT senor is over the calibration standard) is typically made at least once at the start of the build process (after warmup) and typically several times during the build process and at the end of the build process, in order to obtain a traceable NDT sensor calibration dataset associated with a NDT part control dataset and/or with a reference geometry NDT measurement dataset. Advantageously, the calibration standard may be marked with a unique identification code fed into the in-situ and/or remote computing device, either manually or through an automatic code reader. This unique identification code is stored alongside the calibration dataset and the NDT measurement dataset for traceability purposes.

The NDT sensor may be configured to measure the height of the powder and/or the height of sintered portions (i.e. the surface geometry on the trajectory of the NDT sensor). The calibration standard consists of a reference standard with known geometry, typically comprising height steps in the range of 0.01 to 1 mm as calibration features.

In an advantageous embodiment, the NDT sensor configured to measure the height of the last sintered surface is an eddy current sensor. In another advantageous embodiment, the NDT sensor configured to measure this height is an optical sensor.

With the same type of calibration standard, the NDT sensor may be configured to sense a difference between the powder height before sintering and the sintered layer height.

The NDT sensor may be configured to measure the surface roughness of powder that has been spread in the last layer and/or the surface roughness of sintered portions. In this case, a preferred embodiment of the calibration standard consists of a reference standard with known surface roughness and or surface features typically comprising height variations in the range of 0.001 to 0.1 mm as calibration features, installed within +/−1 mm of the plane created by the recoater.

The NDT sensor may be configured to measure the part density. In this case, the calibration standard is a part of known density made of a similar material as the part being produced, and preferably comprising several parts or portions of known density as calibration features, ideally in the density range achievable by the additive manufacturing system.

In an advantageous embodiment, the NDT sensor configured to measure the density of the calibration standard is an eddy current sensor.

The NDT sensor may be configured to measure the powder density prior to solidification. In this case, the calibration standard comprises, as a calibration feature, a portion with a known metallic powder density that is made of metallic powder solidified with a non-conducting, non-magnetic binder material, and preferably several such parts or portions. These parts or portions should ideally have a metallic powder density within or close to the density range achievable with the powder deposition and recoating mechanism of the system

In an advantageous embodiment, the NDT sensor configured to measure the powder density prior to solidification is an eddy current sensor.

100 70 The NDT sensor may be configured to measure defects such as porosity, cracking, lack of fusion or residual stress in the metal pieceand/or in control geometries. In a preferred embodiment, the calibration standard is a piece made of a material similar to that of the printed metal piece, with artificial or natural defects, as calibration features, representative of the defect class searched for. Artificial defects, used as calibration features, may be machined notches to simulate cracks, and/or machined holes to simulate porosity.

26 26 26 a b The NDT sensor may be configured to measure the local electrical conductivity of the metal piece and/or of the control geometries. In this case, the calibration standard is of known electrical conductivity, preferably within +/−50% of the expected sintered part electrical conductivity. The calibration standardmay comprise different portions, for example, a first and a second portion,of different conductivity.

In an advantageous embodiment, the NDT sensor configured to measure the electrical conductivity of the metal piece is an eddy current sensor.

The NDT sensor supporting structure may be the recoater, the print head or another mobile element crossing the build plate whose motorized motion can make the NDT sensor reach the calibration standard.

70 90 16 24 24 16 a FIGS. In an advantageous embodiment, the control geometriescan be designed to represent the most challenging portions for additive manufacturing, for instance a wall or a lattice structureas shown for example in,. They may also be placed to be built on a specific location on the build platformin order to take into consideration an area known to be prone to a given defect category, for instance in a region on the built platformwhere multiple laser beams are close to each other during the additive manufacturing process.

The control geometries can be full cubes or full parallelepipeds, that are suitable for density measurements using Archimedes' method.

The control geometries can be shaped to perform tests of mechanical properties. In particular, the control geometries may exhibit a typical test specimen shaped to be suitable for tensile strength tests, fatigue tests, and/or fatigue crack growth. The fatigue and tensile strength specimens typically exhibit three regions with two cross-sections: The central region, where the failure typically occurs, has a reduced cross-section compared to the extremities which exhibit a larger cross-section. The cross-sectional area, along a plane orthogonal to the longitudinal axis of the specimen, is typically at least 30% larger in the extremities than in the central region. The section typically varies smoothly between the three regions in order to avoid stress concentrations. In the case of tensile strength and fatigue test specimens, the control geometries are also of elongated shape (high aspect ratio), with the central region having a length typically at least five times its width.

70 72 74 72 60 10 13 a a FIGS.to 14 14 a b FIGS., The control geometrycan be shaped to have a closed conductive portionsurrounding an empty spaceas illustrated in particular in. In an embodiment, the closed conductive portionmay be oriented within a plane parallel to a plane within which the recoater or the sensor supporting structure of the additive manufacturing system is configured to move as shown for example inwhere identical control geometries are stacked on top of each other to form a functionalized supporthaving repetitive closed conductive portions at different building layers.

It may however be desirable to have a control geometry with a closed conductive portion extending in different planes, for example to inspect the part quality of an overhang part of a metal piece. The printable threshold overhang angle (PTOA) is the overhang with the lowest angle that can be fabricated without needing support structures while still meeting the part quality requirements. It is defined here as the angle between the XY plane and a surface tangent to the overhang in the XZ (or YZ) plane. Minimizing the PTOA is important, as it reduces the printing and post-processing costs. Documenting that the print quality is achieved for a given PTOA is therefore important.

2 An advantageous way of doing so is to use a closed conductive portion made of several elongated portions with thin sections extending at an angle substantially matching the PTOA (i.e. the minimum overhang angle of a part of the metal piece achievable with the additive manufacturing system without any supporting structure). This PTOA usually varies between 30° to 45° depending on the parameters of the system. Integrating a control geometry on an overhang part may provide valuable information on the properties of the quality on the overhang part and based on this information the angle may be decreased to optimize the additive manufacturing process. The section of thin portions forming the closed control conductive portion is typically between 0.5 and 5 mm.

72 72 90 15 15 a d FIGS.- 16 16 a b FIGS., The closed conductive portionof the control geometry may be oriented in several planes and a closed conductive portion may be formed only upon completion of several build layers. An example of such closed conducting portionis illustrated inand is well adapted to be integrated in a lattice structureas shown into determine the minimum PTOA to be able to optimize the printing process.

70 72 19 19 72 70 72 Such kind of control geometrieswith a closed conductive portionare particularly advantageous to test non-destructively fine structural components with eddy current techniques. In an embodiment of the invention, the NDT sensor is an eddy current sensor comprising a coilgenerating a time-varying electromagnetic field in the range of 1 kHz to 10 MHz. The dimensions of coilin the X and Y directions are substantially equal to the dimensions of the closed conductive portionof the control geometryin the X and Y directions such that the coil may be aligned with the conduction portion.

72 70 In another advantageous embodiment, the NDT sensor is an eddy current sensor comprising a coil generating a time-varying (oscillating) electromagnetic field in the range of 1 kHz to 10 MHz by means of at least one coil with characteristic sensing size equal or close to the dimension of the closed conductive portionof the control geometry.

The characteristic sensing size of an oscillating electromagnetic field produced by the coil around a certain point of interest located within a plane to be investigated (e.g. produced by a generalized coil at a certain distance from a surface) is the distance between the maxima of the amplitude (norm) of the vector potential (A) corresponding to the generated magnetic field (B) along a line belonging to the plane of interest and passing through the point of interest. For this kind of applications (eddy currents), the Coulomb gauge is used to compute the vector potential. In case of an asymmetrical field, the maximum distance over all lines passing through the point of interest is used as the characteristic sensing size.

70 100 70 72 72 19 72 6 FIG. 14 14 a b FIGS.and The control geometrymay be positioned repetitively in lattice-like structures at different heights, either in the metal pieceor in support structures (). Identical control geometrieswith closed conductive portionmay also be stacked on top of and/or next to each other to form a lattice-like structure, as shown in, with repetitive closed conductive portionat different build layers at different heights. The characteristic sensing size of the sensor coilshould be within +/−15%, and preferably within +/−5% of the lattice period defined by the distance in the X and/or Y and/or Z directions between closed conductive portions. These lattice control geometries can be used to test fine structures or lattice structures used to build the part. Lattice structures have the advantage, among other, to reduce the weight of the component while keeping good mechanical properties.

72 70 In an embodiment, the coil is built without ferromagnetic material, and the characteristic sensing size is substantially equal to the size of the coil. In this embodiment, the coil size substantially matches the lattice period and/or the distance between two opposite segments of the closed conductive portionof the control geometry.

The coil may also include magnetic elements affecting the magnetic field distribution in order to shape the magnetic field distribution or achieve electromagnetic shielding between elements. In such a case, the actual size of the coil may differ significantly from its characteristic sensing size.

In another advantageous embodiment, the NDT sensor comprises coils that are spaced apart with a distance matching the lattice period within +/−15%, and preferably within +/−5%.

As obvious to someone skilled in the art, the control geometries may need to be attached to the build plate by means of support structures.

2 FIG. 17 FIG. 27 10 40 27 10 200 40 45 10 The data of the 3D digital model of one or more metal piece which may comprise one or more functionalized supports and control geometries, if any, as described above, are converted by the additive manufacturing software into a build job file. As shown in, the build job file is sent to a control unitof the additive manufacturing system. 3D digital models of metal pieces to be produced are preferably designed on a remote computing deviceas shown inwhich transmits the build job file to the control unitof the additive manufacturing systemthrough the internet. According to an embodiment, the build job file may also be sent from the remote computing deviceto an in-situ computing deviceaccessible by the operator of the systemwhich may modify certain parameters to improve the additive manufacturing process.

100 18 The build job file comprises data to the regions of interest of the metal piecewhich need to be inspected during its additive manufacturing process by NDT sensors.

These data provide enough information on the quality of the metal piece and can be shared without compromising the know-how of the manufacturer of complex mechanical parts as these data reveal only partial information on the internal structure of the metal piece thereby preventing prejudicial reverse-engineering.

In addition, these data are compliant with certifications standards and can therefore be sent to a certified person for certification of AM products in order to advantageously streamline the certification process.

10 Additive manufacturing system 12 Laser source 14 Mirror scanner Sensor supporting structure 16 Reacoater(one embodiment) 18 NDT sensors 19 Coil 20 Material dispensing platform 22 Metallic powder 24 Build platform 24 a Build platform surface 25 Discharge opening 26 Calibration standard 26 a First portion 26 b Second portion 27 Controller 27 a Processor 28 Multiplexer 30 Measurement units 40 Remote computing device 45 In-situ computing device Additive manufacturing software 50 3D model 52 Regions of interest 60 Functionalized support 70 70 70 70 70 70 a b c d e f Control geometries,,,,, 70 Control geometries 72 Closed conductive portion 74 Empty space 80 Empty region 90 Lattice structure 100 Metal piece 102 102 102 102 102 102 102 a b c d e f Regions of interest;,,,,,